Here it is…the finale. Part 3 of the 3-part 802.11ac blog series. We’ve tried to give it to you straight and we hope this has been informative. 802.11ac can be complicated and again, our aim is to tell you what to expect from 802.11ac in the real world and how best to migrate to it as your device population shifts.
So how will 802.11ac Wave 2 improve performance? Beyond the 256-QAM (covered in Part 1) and channel bonding which goes up to 160 MHz channels in wave 2, (covered in Part 2) what else is new in 802.11ac Wave 2, and is it worth the wait? Wave 2 has two additional features: support for 4 spatial streams and an airtime multiplexing feature called MU-MIMO. Let’s explore what use these are in real-world settings.
The 802.11ac standard allows for up to 8 spatial streams. Using a 160 MHz channel and MCS 9 data rates a single radio has a maximum data rate of 6.7 Gbps – another big number being touted by 802.11ac hype-sters. Without a doubt, more spatial streams are a good thing. Like channel bonding, they are another linear performance multiplier, and you have probably already seen the results with 1 stream, 2 stream and 3 stream 802.11n access points.
More streams however, requires more antennas, so just consider what it is going to take to cram 8 Tx and 8 Rx antennas into an AP, and still ensure they perform as well as the antennas in a 3×3 design. Least we mention the increased power requirements. Powering such a beast is sure to exceed the specs for 802.3at, which enterprises are already reluctant to adopt. That’s why with 802.11ac Wave 1, we are seeing a number of 802.11ac AP offerings with only two streams, designed to still operate on legacy 802.3af POE infrastructure. Wave 2 will make the jump to four stream APs, which will undoubtedly require 802.3at POE. Combine this with a 160 MHz channel, and you’re now up to 2.3 Gbps. Not too shabby, if you believe a 160 MHz channel is viable. Again staring reality in the eye, 160 MHz makes sense for mesh and bridging applications, but it’s unlikely to fly where you have high-density client access, and need as many cells as you can manage.
Multi-User MIMO is an exciting development in 802.11ac that will first manifest in Wave 2 products. Normally, with 802.11n and 802.11ac Wave 1, multiple spatial streams are transmitted over multiple antennas to only one client at a time. With each additional stream comes added throughput.
This is all great, when the receiving device is a high-end laptop equipped with multiple receive antennas and can support multiple streams. However, with the explosion of BYOD, the vast majority of mobile devices are not high-end laptops. They are smartphones and tablets which only support one spatial stream. During transmissions to these clients, the other streams go to waste.
MU-MIMO fixes this – well, it half fixes it. It makes it possible to multiplex airtime in such a way that individual clients hear only their part of the message at any given time, and the rest is nulled to them. This allows each of the three streams to be directed to a different one-stream client, simultaneously. So potentially, three clients get serviced in the time it previously took to service one. This has some potential for scaling capacity if done right. The science is not so easy to implement however. When the individual streams are going to different clients, they must be tightly synchronized, so that they don’t create a new type of interference between themselves. At the time of writing, MU-MIMO implementations are still in labs and have not seen the light of day, so it is too early to tell how well it works. But even if it works half as well as it looks on paper, it will boost capacity for all those networks suffering from BYOD overload due to smartphones and tablets. Chip vendor Qualcomm is claiming 2x – 2.5x performance improvements, but that’s hard to believe, because once again there are limitations.
What vendors don’t tell you however, is the fact that the clients cannot reply to the Access Point in MU-MIMO – this is a downlink-only technology that requires a lot of CPU power and resources which low-end clients do not have. Additionally, it is likely that MU-MIMO results in a downshift in encoding scheme, from say 256-QAM to 64-QAM. So it is debatable whether two clients at 64-QAM is better than one at 256-QAM. Another point that is frequently overlooked, is that MU-MIMO uses similar techniques to beamforming, and it is therefore mutually exclusive. The two techniques cannot be used concurrently, so you must consider what you lose by not having beamforming. Once again, what this points to is the need for small cells, with clients close to the AP and ideal RF conditions for MU-MIMO to add value.
For a few extra 802.11ac nuggets, read the full migration guide by clicking below.